'If it still turns on, keep it' feels thrifty - and it is one of the most expensive habits in business IT. An ageing PC does not announce its true cost; it leaks it quietly through lost minutes, support tickets and security gaps. Knowing when to replace a machine, rather than nursing it along, is the difference between planned spend and constant firefighting.
The real cost of keeping old kit
An old computer is rarely free just because it is paid for. Every day it runs slow, it taxes the person using it - a few wasted minutes an hour, multiplied across a year, easily outweighs the cost of a new machine.
On top of that sit the support tickets, the unexpected downtime when something finally gives out, and the productivity hit while a key person waits for a fix. The machine that 'still works' is often the most costly one you own; you just never see the bill in one place.
The signs it is time, not maybe
Some signals are about money and frustration, and once a few stack up the maths has already tipped towards replacement.
- •It is noticeably slow at everyday tasks even after a clean-up - opening apps, switching tabs, saving files.
- •It is generating support tickets or crashing often - each one costs staff and IT time.
- •Repairs are no longer worth it - a costly fix on an old machine is money better put towards a new one.
- •Staff have quietly worked around it - using a phone or another PC because theirs is too painful.
- •It cannot keep up with the software your business now relies on, such as heavier video calls or newer apps.
The signs that are really about security
The other half of the decision is risk, and it is the half firms most often ignore. A computer past its useful life can become a genuine security liability, not just a slow one.
The big trigger is software support ending: when a machine can no longer run a supported, patched operating system, it stops receiving security fixes and becomes a soft target. Windows 11's hardware requirements - including the security chip we cover in what is TPM - left a lot of older PCs unable to upgrade, turning a 'works fine' machine into a liability overnight. An unpatchable PC on your network undermines the basics of endpoint security, and it is one of the first things flagged in a security risk assessment.
A sensible replacement cycle
Rather than waiting for failure, most well-run firms replace on a planned cycle, which turns a nasty surprise into a predictable, budgetable cost. The right interval depends on the machine and the role.
As a rule of thumb: desktops every four to five years, laptops every three to four (they take more knocks and run hotter), and heavy machines like workstations on their own schedule driven by the demands placed on them. Tie the cycle to the budgeting in how much a business PC should cost, and treat it as rolling renewal rather than a one-off cost - the same planned-refresh logic behind the server refresh decision framework.
Repair, redeploy or retire
When a machine reaches the line, you have three sensible options, and the cheapest is not always replacement. Match the choice to the machine's condition and the role.
Repair if it is otherwise healthy and the fix is cheap relative to a new unit. Redeploy a still-decent but underpowered PC to a lighter role - a machine too slow for a power user may be perfect for reception or a meeting room. Retire it properly when it is genuinely past it: that means securely wiping data before disposal or resale, which matters enormously for compliance and is exactly why data handling discipline applies right to the end of a device's life. Done well, replacement is calm, planned and cost-controlled - not a panic when something dies.